ARTIFICIAL NEURAL NETWORK EMBEDDED OPTIMAL MOBILE COMMUNICATION SYSTEM

Abstract

Recent advancement in information and communication has witnessed the involvement and application of embedded system in mobile communications and Internet of Things (IoT) deployment in form of Multiple Operator Enabled SIM (MOES) System. This newly introduced system has shown to have capability to switch from one network to another seamlessly based on Received Signal Strength (RSS). However, such switching capability has shown not to be optimal in nature due to non-application of Artificial Intelligence (AI) in network selection process. Hence, this paper aims at developing Artificial Neural Network (ANN) based system for optimal network selection using RSS. The RSS parameters were collected over time using an existing hardware which reads RSS and the parameters were compiled. The datasets obtained were from four different Mobile Network Operators (MNOs) and used as prediction parameters which were used as input and target parameters for ANN trainings and testing. The parameters were simulated and the results obtained showed high accuracy of 0.99 as the most stable value which means that the developed method was efficient for optimal selection of network as the output regression plots were linear and the graph results of the trainings showed high selection of best RSS values when compared to the actual value results obtained and plotted on the graph. The performance evaluation was carried out by checking the accuracy of the system and the result obtained shows that the system is efficient with 1 as the highest regression value obtained and can be deployed for handing over from one mobile network to another.